Bug 202261

Summary: general protection fault: 0000 [#1] PREEMPT SMP Workqueue: events ttm_bo_delayed_workqueue
Product: Drivers Reporter: Johannes Hirte (johannes.hirte)
Component: Video(DRI - non Intel)Assignee: drivers_video-dri
Status: RESOLVED CODE_FIX    
Severity: normal CC: alexdeucher, hch
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 5.0.0-rc2 Subsystem:
Regression: No Bisected commit-id:

Description Johannes Hirte 2019-01-14 14:22:37 UTC 
first seen with kernel 5.0.0-rc1, now again with 5.0.0-rc2

Jan 14 15:05:49 probook kernel: general protection fault: 0000 [#1] PREEMPT SMP
Jan 14 15:05:50 probook kernel: CPU: 1 PID: 3697 Comm: kworker/1:2 Not tainted 5.0.0-rc2 #66
Jan 14 15:05:50 probook kernel: Hardware name: HP HP ProBook 645 G2/80FE, BIOS N77 Ver. 01.15 03/26/2018
Jan 14 15:05:50 probook kernel: Workqueue: events ttm_bo_delayed_workqueue
Jan 14 15:05:50 probook kernel: RIP: 0010:__memcpy+0x12/0x20
Jan 14 15:05:50 probook kernel: Code: e2 20 48 09 c2 48 31 d3 e9 79 ff ff ff 90 90 90 90 90 90 90 90 90 90 90 0f 1f 44 00 00 48 89 f8 48 89 d1 48 c1 e9 03 83 e2 07 <f3> 48 a5 89 d1 f3 a4 c3 66 0f 1f 44 00 00 48 89 f8 48 89 d1 f3 a4
Jan 14 15:05:50 probook kernel: RSP: 0018:ffffa16a03bd3d78 EFLAGS: 00010246
Jan 14 15:05:50 probook kernel: RAX: bdff99d92e158d48 RBX: 0000000099027000 RCX: 0000000000000200
Jan 14 15:05:50 probook kernel: RDX: 0000000000000000 RSI: ffff967919027000 RDI: bdff99d92e158d48
Jan 14 15:05:50 probook kernel: RBP: ffff967cacd550b0 R08: 0000000000000000 R09: 0000000000000507
Jan 14 15:05:50 probook kernel: R10: 8080808080808080 R11: 0000000000000000 R12: 0000000000001000
Jan 14 15:05:50 probook kernel: R13: 0000000000000000 R14: 0000000000000001 R15: 0000000000000001
Jan 14 15:05:50 probook kernel: FS:  0000000000000000(0000) GS:ffff967caf280000(0000) knlGS:0000000000000000
Jan 14 15:05:50 probook kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Jan 14 15:05:50 probook kernel: CR2: 00007f74431d3000 CR3: 0000000408d6c000 CR4: 00000000001406e0
Jan 14 15:05:50 probook kernel: Call Trace:
Jan 14 15:05:50 probook kernel:  dma_direct_unmap_page+0x68/0x70
Jan 14 15:05:50 probook kernel:  ttm_unmap_and_unpopulate_pages+0xd6/0xf0
Jan 14 15:05:50 probook kernel:  ttm_tt_destroy.part.11+0x44/0x50
Jan 14 15:05:50 probook kernel:  ttm_bo_cleanup_memtype_use+0x29/0x70
Jan 14 15:05:50 probook kernel:  ttm_bo_cleanup_refs+0x10d/0x1d0
Jan 14 15:05:50 probook kernel:  ttm_bo_delayed_delete+0xa3/0x1e0
Jan 14 15:05:50 probook kernel:  ttm_bo_delayed_workqueue+0x12/0x40
Jan 14 15:05:50 probook kernel:  process_one_work+0x1f8/0x420
Jan 14 15:05:50 probook kernel:  worker_thread+0x28/0x3c0
Jan 14 15:05:50 probook kernel:  ? rescuer_thread+0x330/0x330
Jan 14 15:05:50 probook kernel:  kthread+0x10d/0x130
Jan 14 15:05:50 probook kernel:  ? kthread_create_on_node+0x60/0x60
Jan 14 15:05:50 probook kernel:  ret_from_fork+0x22/0x40
Jan 14 15:05:50 probook kernel: Modules linked in: rfcomm bnep joydev kvm_amd btusb ccp btrtl btbcm btintel bluetooth kvm irqbypass hp_wmi wmi_bmof ecdh_generic uvcvideo crc32_pclmul videobuf2_vmalloc iwlmvm videobuf2_memops videobuf2_v4l2 videode
v mac80211 videobuf2_common snd_hda_codec_conexant hid_logitech_hidpp snd_hda_codec_generic aesni_intel snd_hda_codec_hdmi aes_x86_64 snd_hda_intel crypto_simd cryptd snd_hda_codec glue_helper iwlwifi psmouse snd_hwdep i2c_piix4 k10temp fam15h_pow
er snd_hda_core snd_pcm cfg80211 realtek snd_timer rfkill snd r8169 soundcore libphy wmi hp_wireless hid_logitech_dj rtsx_pci_sdmmc mmc_core xhci_pci ehci_pci xhci_hcd ehci_hcd rtsx_pci efivarfs autofs4
Jan 14 15:05:50 probook kernel: ---[ end trace 06764c6abeba2453 ]---
Comment 1 Johannes Hirte 2019-01-14 14:24:25 UTC 
Hardware is an AMD PRO A10-8700B R6 CARRIZO
Comment 2 Alex Deucher 2019-01-14 14:25:58 UTC 
Please attach your dmesg output and xorg log (if using X).  Is there a trigger for this or is it random?
Comment 4 Johannes Hirte 2019-01-14 21:10:29 UTC 
(In reply to Michel Dänzer from comment #3)
> Looks like https://bugs.freedesktop.org/109234 =>
> https://lists.freedesktop.org/archives/dri-devel/2019-January/202909.html
> and follow-ups.

Yes, looks like it's this bug.

From https://lists.linuxfoundation.org/pipermail/iommu/2019-January/032651.html

> Hmm, I wonder if we are not actually using swiotlb in the end,
> can you check if your dmesg contains this line or not?
>
> PCI-DMA: Using software bounce buffering for IO (SWIOTLB)

No, it isn't. I can't find anything about SWIOTLB or PCI-DMA in dmesg.

Just testing the patch. I don't have a reproducer, but it could be suspend/resume related.
Comment 5 Johannes Hirte 2019-01-16 18:12:58 UTC 
With the change from https://lists.linuxfoundation.org/pipermail/iommu/2019-January/032651.html see bug seems to be fixed. I've not encountered any Oops since testing with this change.
Comment 6 Johannes Hirte 2019-01-18 20:32:03 UTC 
fixed with commit 227a76b64718888c1687cc237463aa000ae6fb2b